Question 1
1 markA hospital compared the waiting times for appointments at two of its clinics. At one clinic, a patient waited 94 days because his appointment was cancelled twice. All the other waiting times at the two clinics were between 3 and 41 days.
Give one reason why the interdecile range is better than the range for comparing the spread of the waiting times at the two clinics.
Model answer
The range uses only the shortest and the longest waits, so the one wait of 94 days would make the range for that clinic much bigger. The interdecile range (90th percentile - 10th percentile) covers the middle 80% of the waiting times, so it is not affected by this extreme value.
Explanation
Measures of spread based on percentiles or quartiles, such as the interdecile and interquartile ranges, leave out the extreme ends of the data, so an outlier does not distort them. The range depends only on the two most extreme values, so one unusual wait can make a comparison of ranges misleading.
Mark scheme
-
1 mark
B1 for the interdecile range is not affected by extreme values such as the wait of 94 days (it uses only the middle 80% of the data), but the range is
oe, e.g. 'the range uses only the shortest and longest waits, so the one wait of 94 days makes that clinic's range much bigger'. Accept 'outlier' for extreme value. B0 for 'it is more accurate', 'it is more reliable' or 'it uses more of the data' with no reference to extreme values or outliers.